On Fri, 25 Mar 2016 13:32:50 +0200 Nerijus Baliunas <neri...@users.sourceforge.net> wrote:
NB> I have some old messages, and when I try to open them, dovecot dies: You probably should report this to Dovecot people in any case... NB> I get an assert in M: NB> NB> Program received signal SIGTRAP, Trace/breakpoint trap. NB> 0x00000000005f4b01 in MessageCC::GetBody (this=0xfc0a10) at .src/mail/MessageCC.cpp:738 NB> 738 ASSERT_MSG( m_Body && m_Envelope, _T("failed to get body and envelope!") ); NB> (gdb) bt NB> #0 0x00000000005f4b01 in MessageCC::GetBody() (this=0xfc0a10) at .src/mail/MessageCC.cpp:738 NB> #1 0x00000000005f5467 in MessageCC::CheckBody() const (this=0xfc0a10) at /a/M/M/include/MessageCC.h:176 NB> #2 0x00000000005f3669 in MessageCC::CheckMIME() const (this=0xfc0a10) at .src/mail/MessageCC.cpp:507 NB> #3 0x00000000005f374c in MessageCC::GetTopMimePart() const (this=0xfc0a10) at .src/mail/MessageCC.cpp:532 NB> #4 0x000000000048c2ac in MessageView::Update() (this=0xee6400) at .src/classes/MessageView.cpp:3060 The assert in GetBody() definitely seems wrong, there doesn't seem any reason to assume that mail_fetchstructure() can't fail -- and we even see that it can. So it should be removed and replaced with an ERRORMESSAGE(). NB> Usually after a few continues M crashes. This must be due to dereferencing either the envelope or body pointer somewhere without checking if it's null, but I don't immediately see where does happen. Can you catch the crash under the debugger as well please? Thanks, VZ
pgp6QzrxKv5W8.pgp
Description: PGP signature
------------------------------------------------------------------------------ Transform Data into Opportunity. Accelerate data analysis in your applications with Intel Data Analytics Acceleration Library. Click to learn more. http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
_______________________________________________ Mahogany-Developers mailing list Mahogany-Developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mahogany-developers